Many highly proficient bilinguals engage in spontaneous code-switching, fluidly alternating between their two languages when conversing with each other. Experimental methods that have investigated language switching in bilinguals consistently reveal costs (i.e. longer naming times) to switching. The experimental evidence stands at odds with the seamless integration and frequency of use of both languages in spontaneous code-switching. This proposal aims to investigate the cognitive and neural processes that support the ability to integrate both languages during speaking and listening. In so doing, we will develop a new fMRI method that makes use a referential communication task with an experimenter-confederate. A developmental phase with English monolinguals will refine and optimize the methodological parameters and analyses of the new method. Experiment 1 will use this new conversational paradigm to test Spanish-English bilinguals who are also members of the U.S. Latino population. Bilingual subjects will complete the referential communication task with a bilingual experimenter from the same population thus setting an appropriate conversation context for spontaneous code-switching. Analyses will investigate the neural correlates of the production and comprehension of spontaneous code-switching. In Experiment 2, the same bilingual individuals will complete a cued language switching task also with fMRI. The results of Experiments 1 and 2 will be analyzed to compare the neural overlap between spontaneous code-switching and cued language switching. The development of a new neuroimaging technique that focuses on spontaneous conversation in bilinguals will significantly contribute to the methodological tools available to language researchers generally, and more importantly, it will broaden the research focus to minority language and dialect speech acts that have been underrepresented in the language sciences.